Extensible surface for consuming information extraction services

• US 9,064,004 B2
• Filed: 03/04/2011
• Issued: 06/23/2015
• Est. Priority Date: 03/04/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: DeRose, Pedro Dantas
• Primary Examiner(s): BOWEN, RICHARD L
• Application Number:
  US13/040,939
• Publication Number:
  US 20120226715A1
• Time in Patent Office: 1,572 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/254 Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses
• CPC Class Codes: G06F 16/284 Relational databases
• CPC Class Codes: G06F 16/30 of unstructured textual data document management systems G06F16/93
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

Representing structured data extracted from unstructured data in fashion allowing querying using relational database concepts. A method includes receiving user input specifying one or more database views. The method further includes receiving user input specifying an information extraction technique, such as an extraction workflow. The method further includes receiving user input specifying a corpus of data. The extraction technique is applied to the corpus of data to produce the one or more database views. These views can then be queried or operated on using database tools.

Claims

1. In a computing environment, a method of representing structured data extracted from unstructured data in a fashion which allows querying using relational database concepts, the method comprising:

receiving user input specifying one or more database views;

receiving user input specifying an information extraction technique, the information extraction technique defining how to extract structured data from unstructured data and the information extraction technique comprising a phrase semantic extraction technique which determines a semantic relationship about one or more words based upon a semantic environment of the one or more words;

receiving user input specifying a corpus of data comprising unstructured data, the unstructured data comprising data that is not organized semantically such that it does not have a formalized type and is not in a formal entity level relationship; and

applying the extraction technique to the corpus of data to extract structured data from the unstructured data of the corpus of data and to produce the one or more database views including the extracted structured data.

2. The method of claim 1, wherein the one or more views comprises one or more ad hoc views for a single query.

3. The method of claim 1, wherein multiple views are specified as a collection of views.

4. The method of claim 1, further comprising receiving additional options related to the extraction technique.

5. The method of claim 1, further comprising performing one or more queries on the produced views.

6. The method of claim 1, wherein applying the extraction technique further produces procedures.

7. The method of claim 6, wherein the procedures define methods of operating on the one or more views.

8. The method of claim 1, wherein the extraction technique comprises a property promotion.

9. The method of claim 1, wherein the extraction technique comprises a relationship identification.

10. The method of claim 1, wherein the extraction technique comprises an entity recognition or entity extraction.

11. The method of claim 1, wherein the extraction technique comprises entity disambiguation.

12. The method of claim 1, wherein the one or more views represent a graph.

13. The method of claim 1, wherein the one or more views represent tabular data.

14. The method of claim 1, wherein the extraction technique comprises pattern recognition.

15. One or more physical computer readable storage media comprising computer executable instructions that when executed by one or more processors cause the following to be performed:

accept user input specifying one or more database views;

accept user input specifying an information extraction technique, the information extraction technique defining how to extract structured data from unstructured data and the information extraction technique comprising a phrase semantic extraction technique which determines a semantic relationship about one or more words based upon a semantic environment of the one or more words;

accept user input specifying a corpus of data comprising unstructured data, the unstructured data comprising data that is not organized semantically such that it does not have a formalized type and is not in a formal entity level relationship; and

apply the extraction technique to the corpus of data to extract structured data from the unstructured data of the corpus of data and to produce the one or more database views including the extracted structured data.

16. The one or more physical computer readable storage media of claim 15, wherein the one or more or more views comprises one or more ad hoc views for a single query.

17. The one or more physical computer readable storage media of claim 15, wherein the one or more views is a single view that is itself a collection of views.

18. The one or more physical computer readable storage media of claim 15, wherein applying the extraction technique further produces procedures.

19. A system for representing structured data extracted from unstructured data in a fashion which allows querying using relational database concepts, the system comprising one or more processors and one or more physical computer readable storage media having encoded thereon computer executable instructions which, when executed upon the one or more processors, cause the system to:

store one or more corpuses of data, each of the one or more corpuses of data comprising unstructured data;

receive user input specifying one or more database views;

receive user input specifying an information extraction technique, the information extraction technique defining how to extract structured data from unstructured data and the information extraction technique comprising a phrase semantic extraction technique which determines a semantic relationship about one or more words based upon a semantic environment of the one or more words;

receive user input specifying one or more of the stored one or more corpuses of data comprising unstructured data, the unstructured data comprising data that is not organized semantically such that it does not have a formalized type and is not in a formal entity level relationship; and

apply the extraction technique to the specified one or more corpuses of data to extract structured data from the unstructured data of the corpuses of data and to produce the one or more database views including the extracted structured data.